A Land Use and Land Cover Classification System for Use with Remote Sensor Data

GEOLOGICAL SURVEY PROFESSIONAL PAPER 964

A Land Use and Land Cover Classification System for Use with Remote Sensor Data

By JAMES R. ANDERSON, ERNEST E. HARDY, JOHN T. ROACH, and RICHARD E. WITMER

GEOLOGICAL SURVEY PROFESSIONAL PAPER 964

A revision of the land use classification system as presented in U.S. Geological Survey Circular 671

UNITED STATES GOVERNMENT PRINTING OFFICE, vVASHINGTON 1976 UNITED STATES DEPARTMENT OF THE INTERIOR

JAMES G. WATT, Secretary

GEOLOGICAL SURVEY

Dallas L. Peck, Director

First printing 1976 Second printing 1976 Third printing 1978 Fourth printing 1983

For sale by the Distribution Branch, U.S. Geological Survey, 604 South Pickett Street, Alexandria, VA 22304 CONTENTS

Page Abstract 1 Introduction 1 ~eed for standardization ------­ 2 Historical development of the classification system ------3 Designing a classification system for use with remote :sensing techniques ______4 Classification criteria ______------______5 Developing the classification system ------7 Using the classification system ______9 Definitions ------10 Urban or Built-up Land ------10 Agricultural Land ------13 14 RangruandForest Land ------______16 VVater ------­ 17 VVetland ------­ 17 Barren Land ------­ 18 Tundra ------20 Perennial Snow or Ice ------21 Map presentation ------____ ------22 Selected bibliography ------.... ------27

ILLUSTRATIONS

FIGURE 1. Map of ,a part of the Indianapolis, Ind.-111., 1:250,000 quadrangle, showing Level I land use and land cover ------23 2. Map of a part of the Indianapolis, Ind.-111., 1:250,000 quadrangle, showing Level II land use and land cover ------24 3. Map of a part of the Maywood, Ind., 1:24,000 quadrangle, showing Level II land use and land cover __ 25 4. Map of a part of the Maywood, Ind., 1:24,000 quadrangle, showing Level III land use and land cover __ 26

TABLES

TABLE 1. Major uses of land, United S.tates, 1969 ------·------8 2~ Land use and land cover classification system for use with remote sensor data ------8 3. Standard land use code-first level categories ------·------9 4. U.S.G.S. Level I land use color code ------22

III

A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA

By JAMEs R. ANDERSON, ERNEST E. HARDY, JoHN T. RoAcH, and RICHARD E. WITMER

ABSTRACT (296,000 hectares) were urbanized each year, trans­ The framework of a national land use and land cover portation land uses expanded by 130,000 acres classification system is presented for use with remote sensor (53,000 hectares) per year, and recreational area data. The classification system has been developed to meet increased by about 1 million acres ( 409,000 hec­ the needs of Federal and State agencies for an up-to-date tares) per year. Knowledge of the present distribu­ overview of land use and land cover throughout the country on a basis that is uniform in categorization at the more tion and area of· such agricultural, recreational, and generalized first and second levels and that will be receptive urban lands, as well as information on their chang­ to data from satellite and aircraft remote sensors. The pro­ ing proportions, is needed by legislators, planners, posed system uses the features of ·existing widely used classi­ and State and local governmental officials to deter­ fication systems that are amenable to data derived from re­ mine better land use policy, to project transporta­ mote sensing sources. It is intentionally l·eft open-ended so that Federal, regional, State, and local agencies can have tion and utility demand, to identify future develop'"" flexibility in developing more detailed land use classifications ment pressure points and areas, and to implement at the third and fourth levels in order to meet their particular effective plans for regional development. As Claw­ needs and at the ~arne time rema:in compatible with each son and Stewart ( 1965) have stated: other and the national system. Revision of the land use classification system as presented in U.S. Geological Survey In this dynamic situation, accurate, meaningful, current Circular 671 was undertaken in order to incorporate the re­ data on land use are essential. If public agencies and private sults of extensive testing and review of the categorization organizations are to know what is happening, and are to make and definitions. sound plans for their own futwre action, then reliable infor­ mation is critical. INTRODUCTION The variety of land use and land cover data needs is exceedingly broad. Current land use and land cov­ A modern nation, as a modern business, must have er data are needed for equalization of tax assess­ adequate information on many complex interrelated ments in many States. Land use and land cover data aspects of its activities in order to make decisions. also are needed by Federal, State, and local agencies Land use is only one such aspect, but knowledge for water-resource inventory, flood control, water­ about land ~se and land cover has become increas­ supply planning, and waste-water treatment. Many ingly important as the Nation plans to overcome Federal agencies need current comprehensive inven­ the problems of haphazard, uncontrolled develop­ tories of existing activities on public lands combined ment, deteriorating environmental quality, loss of with the existing and changing uses of adjacent prime agricultural lands, destruction of important private lands to improve the management of public wetlands, and loss of fish and wildlife habitat. Land lands. Federal agencies also need land use data to use data are needed in the analysis of environmental assess the environmental impact resulting from the processes and problems that must be understood if development of energy resources, to manage wildlife living conditions and standards are to be improved or maintained at current levels. resources and minimize man-wildlife ecosystem One of the prime prerequisites for better use of conflicts, to make national summaries of land use land is information on existing land use patterns patterns and changes for national policy formula.. and changes in land use through time. The U.S. tion, and to prepare environmental impact state­ Department of Agriculture (1972) reported that ments and assess future impacts on environmental during the decade of the 1960's, 730,000 acres quality.

1 2 A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA

NEED FOR STANDARDIZATION which land use and land cover types and their bound­ aries are interpreted. During the past 40 years For many years, agencies at the various govern­ several surveys, studies, and other projects have mental levels have been collecting data about land, successfully demonstrated that remote sensor data but for the most part they have worked independent­ are useful for land lise and land cover inventory and ly and without coordination. Too often this has mapping. These surveys have contributed to our con­ meant duplication of effort, or it has been found that fidence that land use and land cover surveys of larger data collected for a specific purpose were of little areas are possible by the use of remote sensor data or no value for a similar purpose only a short time later. bases. In the mid-1940's, Francis J. Marschner began There are many different sources of information mapping major land use associations for the entire on existing land use and land cover and on change§ United States, using aerial photographs taken dur­ that are occurring. Local· planning agencies make ing the late 1930's and the early 1940's.. Marschner use of detailed information generated during ground produced a set of State land use maps at the scale of surveys involving enumeration and observation. In­ 1 : 1,000,000 from mosaics of the aerial photographs terpretation of large-scale aerial photographs also and then compiled a map of major land uses at has been used widely (Avery, 1968). In some cases, 1:5,000,000 (Marschner, 1950). supplementary information is inferred on the basis More recently, the States of New York and Min­ of utility hookups, building permits, and similar in­ nesota have used remote sensor data for statewide formation. Major problems are present in the appli­ land use mapping. New York's LUNR (Land Use cation and interpretation of the existing data. These and Natural Resources) Program (New York State include changes in definitions of categories and data­ Office of Planning Coordination, 1969) employs com­ collection methods by source agencies, incomplete puter storage of some 50 categories of land use infor­ data coverage, varying data age, and employment of mation derived from hand-drafted maps compiled by incompatible classification systems. In addition, it interpreting 1967-1970 aerial photography. This is nearly impossible to aggregate the available data information can be updated and manipulated to pro­ because of the differing classification systems used. vide numerical summaries and analyses and com­ The demand for standardized land use and land puter-generated maps (Hardy and Shelton, 1970). cover data can only increase as we seek to assess Aerial photographs taken in the spring of 1968 and and manage areas of critical concern for environ­ 1969 at an altitude of about 40,000 ft ( 12,400 m) mental control such as flood plains and wetlands, yielded the data incorporated into the nine categories energy resource development and production areas, of the Minnesota Land Use Map, a part of the Min­ wildlife habitat, recreational lands, and areas such nesota Land Management Information System (Or­ as major residential and industrial development sites. ning and Maki, 1972). Thrower's map (1970) of the As the result of long concern about duplication Southwestern United States represents the first and coordination among Federal, State, and local large-area inventory of land use employing satellite governments in the collection and handling of vari­ imagery. Imagery from several manned and unman­ ous types of data, the United States has already ned missions was used in deriving the general land achieved reasonably effective, though not perfect, use map published at a scale of 1: 1,000,000. standardization in some instances, as evidenced by Remote sensing techniques, including the use of present programs in soil surveys, topographic map­ conventional aerial photography, can be used effec­ ping, collection of weather information, and inven­ tively to complement surveys based on ground ob­ tory of forest resources. Recent developments in servation and enumeration, so the potential of a data processing and remote sensing technology make timely and accurate inventory of the current use of the need for similar cooperation in land use inven­ the Nation's land resources now exists. At the same tories even more evident and more pressing. Devel­ time, data processing techniques permit the storage opment and acceptance of a system for classifying of large quantities of detailed data that can be or­ land use data obtained primarily by use of remote ganized in a variety of ways to meet specific needs. sensing techniques, but reasonably compatible with The patterns of resource use and resource demand existing classification systems, are the urgently are constantly changing. Fortunately, the capability needed first steps. to obtain data about land uses related to resource This is not the first time that use of remote sensors development is improving because of recent tech­ has been proposed to provide the primary data from nological improvements in remote sensing equip- HISTORICAL DEVELOPMENT OF THE CLASSIFICATION SYSTEM 3

ment, interpretation techniques, and data process­ (1965), the inventory of Major Uses of Land made ing (National Academy of Sciences, 1970). every 5 years by the Economic Research Service of the U.S. Department of Agriculture (Frey, 1973), HISTORICAL DEVELOPMENT OF THE and the national inventory of soil and water conser­ CLASSIFICATION SYSTEM vation needs, initiated in 1956 and carried out \for the second time in 1966 by several agencies of the The needs of Federal agencies for a broad over­ U.S. Departmentsof Agriculture and Interior (U.S. view of national land use and land cover patterns Department of Agriculture, 1971). and trends and environmental values led to the for­ Two land use classification systems initially pro­ mation of an Interagency Steering Committee on posed by James R. Anderson for conference use were Land Use Information and Classification early in designed to place major reliance· on remote sensing, 1971. The work of the committee, composed of rep­ although supplementary sources of information were resentatives from the Geological Survey of the U.S. assumed to be available for the more elaborate of the Department of the Interior, the National Aero­ two (Anderson, 1971). The classification system for nautics and Space Administration (NASA), the the New York State Land Use and Natural Re­ Soil Conservation Service of the U.S. Department of sources Inventory, developed mainly at the Center Agriculture, the Association of American Geograph­ for Aerial Photographic Studies at Cornell Univer­ ers, and the International Geographical Union, has sity, had been designed for use with aerial photogra­ been supported by NASA and the Department of the phy at 1 : 24,000 scale, and although devised speoifi­ Interior and coordinated by the U.S. Geological cally for New York State, it was adaptable for rlse Survey (U.S.G.S.). elsewhere. To take advantage of the New York The objective of the committee was the develop­ experience, Ernest E. Hardy and John T. Roach ment of a national classification system that would were invited to collaborate in preparing the definitive be receptive to inputs of data· from both convention­ framework of the proposed classification. Definitions al sources and remote sensors on high-altitude air­ of land use categories used in New York were care­ craft and satellite platforms, and that would at the fully reviewed and were modified to make them ap­ same time form the framework into which the cate­ plicable to the country as a whole. The resulting gories of more detailed land use studies by regional, classification was presented in U.S. Geological Sur­ State, and local agencies could be fitted and aggre­ vey Circular 671. Because of his past experience with gated upward from Level IV toward Level I for the Commission on Geographic Applications of more generalized smaller scale use at the national Remote Sensing of the Association of American Ge­ level. ographers, Richard E. Witmer was invited to partici­ pate with the others in this revision of the classifica­ Several classification systems designed for or tion system. amenable to use with remote sensing techniques served as the basis for discussion at a Conference on Attention was given mainly to the more general­ Land Use Information and Classification in Wash­ ized first and second levels of categorization. Defini­ ington~ D.C., June 28-30, 1971. This conference was tions for each of the categories on these two levels attended by more than 150 representatives of Fed­ were subjected to selective testing and evaluation by eral, State, and local government agencies, univer­ the U.S.G.S., using dalta obtained primarily from sities, institutes, and private concerns. On the basis high-altitude flights as part of the research in con­ of these discussions, the Interagency Steering Com­ nection with the U.S.G.S. Central Atlantic Regional mittee then proposed to develop and test a land use Ecological Test Site (CARETS) Project (28,800 2 2 and land cover classification system that could be mi or 74,700 km ), the Phoenix Pilot Project 2 2 used with remote sensing and with minimal rei'iance (31,500 mi or 81,500 km ), and the land use mapping on supplemental information at the more generalized for the Ozarks Regional Commission (72,000 miZ or 2 first and second levels of categorization. The need 186,500 km ). for compatibility with the more generalized levels of The work of Pettinger and Poulton ( 1970) pro­ land use and land cover categorization in classifica­ vided valuable insight into the land use mosaic of the tion systems currently in use was clearly recognized, Southwestern United States. Some of the categoriza­ especially those levels of the Standard Land Use Cod­ tion for barren land and rangeland suggested by ing Manual published by the U.S. Urban Renewal these researchers has been adopted in this land use Administration and the Bureau of Public Roads and land cover classification system. 4 A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA

DESIGNING A CLASSIFICATION SYSTEM·· FOR particularly amenable to interpretation from remote USE WITH REMOTE SENSING TECHNIQUES sensor data. For example, hunting is a very common and pervasive recreational use of land, but hunting There is no one ideal classification of land use and usually occurs on land that would be classified as land cover, and it is unlikely that one could ever be some type of forest, range, or agricultural land developed. There are different perspectives in the either during ground survey or image interpretation. classification process, and the process itself tends to Consequently, supplemental information is needed be subjective, even when an objective numerical ap­ to identify lands used for hunting. Supplemental in­ proach is used. There is, in fact, no logical reason to formation such as land ownership maps also is neces­ expect that one detailed inventory should be adequate sary to determine the use of lands such as parks, for more than a short time, since land use and land game refuges, or water-conservation districts, which cover patterns change in keeping with demands for may have land uses coincident with administrative natural resources. Each classification is made to suit boundaries not usually discernable by inventory the needs of the user, and few users will be satisfied using remote sensor data. For these reasons, types of with an inventory that does not meet most of their land use and land cover identifiable primarily from needs. In attempting to develop a classification sys­ remote sensor data are used as the basis for organiz­ tem for use with remote sensing techniques that will ing this classification system. Agencies requiring provide a framework to satisfy the needs of the more detailed land use information may need to majority of users, certain guidelines of criteria for employ more supplemental data. evaluation must first he established. In almost any classification process, it is rare to To begin with, there is considerable diversity of find the clearly defined classes that one would like. opinion about what constitutes land use, although In determining land cover, it would seem simple to present use of land is one of the characteristics that draw the line between land and water until one con­ is widely recognized as significant for planning and siders such problems as seasonally wet areas, tidal management purposes. One concept that has much flats, or marshes with various kinds of plant cover. merit is that land use refers to, "man's activities on Decisions that may seem arbitrary must be made at land which are directly related to the land" (Claw­ times, but if the descriptions of categories are com­ son and Stewart, 1965). Land cover, on the other plete and guidelines are explained, the inventory hand, describes, "the vegetational and artificial con­ process can be repeated. The classification system structions covering the land surface" (Burley, must allow for the inclusion of all parts of the area 1961), under study and should also provide a unit of refer­ The types of land use and land cover categoriza­ ence for each land use and land cove:r type. tion developed in the classification system presented The problem of inventorying and classifying in this report can be related to systems for classify­ multiple uses occurring on a single parcel of land ing land capability, vulnerability to certain manage­ will not be easily solved. Multiple uses may occur ment practices, and potential for any particular ac­ simultaneously, as in the instance of agricultural tivity or land value, either intrinsic or speculative. land or forest land used for recreational activities Concepts concerning lancf cover and land use ac­ such as hunting or camping. Uses may also occur tivity are closely related and in many cases have alternately, such as a major reservoir providing been used interchangeably. The purposes for which flood control during spring runoff and generating lands are being used commonly have associated types power during winter peak demand periods. This of cover, whether they be forest, agricultural, resi­ same reservoir may have sufficient water depth to be dential, or industrial. Remote sensing image-form­ navigable by commercial shipping the year round ing devices do not record activity directly. The and may additionally provide summer recreational remote sensor acquires a response which is based on opportunities. Obviously all of the~e activities would many characteristics of the land surface, including not be detectable on a single aerial photograph. How­ natural or artificial cover. The interpreter uses pat­ ever, interpreters have occasionally related flood­ terns, tones, textures, shapes, and site associations control activities to drawdown easements around to derive information about land use activities from reservoirs detectable on imagery acquired during what is basically information about land cover. winter low-water levels. Similarly, major locks at Some activities of man, however, cannot be direct­ water-control structures imply barge or ship traffic, ly related to the type of land cover. Extensive recrea­ and foaming tailraces indicate power generation. tional activities covering large tracts of land are not Pleasure-boat marinas, as well as the wakes of the CLASSIFICATION CRITERIA 5 boats themselves, can be detected on high-altitude 2. The accuracy of interpretation for the several photographs. Although each of these activities is categories should be about equal. detectable at some time using remote sensing, many 3. Repeatable or repetitive results should be ob­ other multiple-use situations cannot be interpreted tainable from one interpreter to another and with the .same degree of success. The example of the from one time of sensing to another. reservoir does provide insight into another facet of 4. The classification system should be applicable the problem's solution, however, and that is the pos­ over extensive areas. sibility and need for acquiring collateral data to aid 5. The categorization should permit vegetation in the understanding of a multiple-use situation. and other types of land cover to be used as sur­ The vertical arrangement of many uses above and rogates for activity. below the actual ground surface provides additional 6. The classification system should be suitable for problems for the land use interpreter. Coal and use with remote sensor data obtained at differ­ other mineral deposits under croplands or forests, ent times of the year. electrical transmission lines crossing pastures, ga.­ 7. Effective use of subcategories that can be ob­ rages underground or on roofs of buildings, and sub­ tained from ground surveys or from the use of ways beneath urban areas all exemplify situations larger scale or enhanced remote sensor data which must be resolved by individual users and com­ should be possible. pilers of land use data. 8. Aggregation of categories must be possible. The size of the minimum area which can be de­ 9. Comparison with future land use data should picted as being in any particular land use category be possible. depends partially on the scale and resolution of the 10. Multiple uses of land should be recognized when original remote sensor data or other data source possible. from which the land use is identified and interpreted. Some of these criteria should apply to land use It also depends on the scale of data compilation as and land cover classification in general, but some of well as the final scale of the presentation of the the criteria apply primarily to land use and land land use information. In some cases, land uses can­ cover data interpreted from remote sensor data. not be identified with the level of accuracy approach­ ing the size of the smallest unit mappable, while in It is hoped that, at the more generalized first and others, specific land uses can be identified which are second levels, an accuracy in interpretation can be too small to be mapped. Farmsteads, for example, attained that will make the land use and land cover are usually not distinguished from other agricultural data comparable in quality to those obtained in other land uses when mapping at the more generalized ways. For land use and land cover data needed for levels of the classification. On the other hand, these planning and management purposes, the accuracy of farmsteads may well be interpretable but too small interpretation at the generalized first and second to be represented at the final format scale. Analogous levels is satisfactory when the interpreter makes the situations may arise in the use of other categories. correct interpretation 85 to 90 percent of the time. When maps are intended as the format for pre­ For regulation of land use activities or for'tax assess­ senting land use data, it is difficult to represent any ment purposes, for example, greater accuracy usual­ unit area smaller than 0.10 inch (2.54 mm) on a side. ly will be required. Greater accuracy generally will In addition, smaller areas cause legibility problems be attained only at much higher cost. The accuracy for the map reader. Users of computer-generated of land use data obtained from remote sensor sources graphics are similarly constrained by the minimum is comparable to that acquired by using enumeration size of the computer printout. techniques. For example, post.enumeration surveys made by the U.S. Bureau of the Census revealed that CLASSIFICATION CRITERIA 14 percent of all farms (but not necessarily 14 per­ A land use and land cover classification system cent of the farmland) were not enumerated during which can effectively employ orbital and high-alti­ the 1969 Census of Agriculture (Ingram and Pro­ tude remote sensor data should meet the following chaska, 1972). criteria (Anderson, 1971) : In addition to perfecting new interpretation tech­ 1. The minimum level of interpretation accuracy niques and procedures for analysis, such as the vari­ in the identification of land use and land cover ous types of image enhancement and signature iden­ categories from remote sensor data should be at tification, we can assume that the resolution capa­ least 85 percent. bility of the various remote sensing systems will also 6 A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA improve. Resolution, or resolving power, of an imag­ and generate local information at the intrastate, re­ ing system refers to its ability to separate two gional, county, or municipal level. It is intended that objects some distance apart. In most land use appli­ these latter levels of categorization will be developed cations, we are most interested in the minimum size by the user groups themselves, so that their specific of an area which can be recognized as having an needs may be satisfied by the categories they intro­ interpretable land use or land cover type. Obviously, duce into the structure. Being able to aggregate more such a minimum area depends not only on the type detailed categories into the categories at Level II and characteristics of the imaging system involved, being adopted by the U.S.G.S. is desirable if the but pragmatically also on the order of "generation" classification system is to be useful. In general,Level of the imagery, that is, how far the study image is II land use and land cover data interface quite effec­ removed in number of reproduction stages from the tively with point and line data available on the stand­ original record. The user should refer to the most ard U.S.G.S. topographic maps. recent information available in determining the reso­ This general relationship between the categoriza­ lution parameters of the system. tion level and the data source is not intended to The kind and amount of land use and land cover restrict users to particular scales, either in the information that may be obtained from different original data source from which the land use infor­ sensors depend on the altitude and the resolution of mation is compiled or in the final map product or each sensor. There is little likelihood that any one other graphic device. Level I land use information, sensor or system will produce good data at all alti­ for example, while efficiently and economically gath­ tudes. It would be desirable to evaluate each source ered over large areas by a LANDSAT type of satel­ of remote sensing data and its application solely on lite or from high-altitude imagery, could also be the basis of the qualities and characteristics of the interpreted from conventional large-scale aircraft source. However, it is common practice to transfer imagery or compiled by ground survey. This same the data to a base map, and no matter what the information can be displayed at a wide variety of guidelines are, it is difficult to use a base map with­ scales ranging from a standard topographic map out extracting some additional data from such maps. scale, such as 1 : 24,000 or even larger, to the much Topographic maps, road maps, and detailed city smaller scale of the orbital imagery, such as maps will generally contribute detail beyond the 1: 1,000,000. Similarly, several Level II categories capabilities of the remote sensor data. (and, in some instances, Level III categories) have The multilevel land use and land cover classifica­ been interpreted from LANDSAT data. Presently, tion system described in this report has been devel­ though, Level II categories are obtained more accur­ oped because different sensors will provide data at a ately from high-altitude photographs. Much Level range of resolutions dependent upon altitude and III and Level IV land use and land cover data can scale. In general, the following relations pertain, also be obtained from high-altitude imagery. This assuming a 6-inch focal length camera is used in level of categorization can also be presented at a obtaining aircraft imagery. wide range of scales. However, as the more detailed levels of categorization are used, more dependence Classification level Typical data characteristics necessarily must be placed on higher resolution re­ I ------LANDSAT (formery ERTS) type of data. mote sensor data and supplemental ground surveys. II ______High-altitude data at 40,000 ft (12,400 m) The principal remote sensor source for Level II or above (less than 1:80,000 scale). III ,______Medium-altitude data taken between 10,000 data at the present time is high-altitude, color-infra­ and 40,000 ft (3,100 and 12,400 m) red photography. Scales smaller than 1:80,000 are ( 1:20,000 to 1:80,000 scale). IV ______Low-altitude data taken below 10,000 ft characteristic of high-altitude photographs, but (3,100 m) (more than 1:2'0,000 scale). scales from 1 :24,000 to 1 : 250,000 generally have Although land use data obtained at any level of been used for the final map products. categorization certainly should not be restricted to The same photography which now is used to con­ any particular level of user groups nor to any par­ struct or update 1 : 24,000 topographic maps or ortho­ ticular scale of presentation, information at Levels photoquads at similar scales is a potential data I and II would generally be of interest to users who source for inventorying land use and land cover. The desire data on a nationwide, interstate, or statewide orthophoto base, in particular, commonly can enable basis. More detailed land use and land cover data rapid interpretation of Levels I and II informa­ such as those categorized at Levels III and IV usual­ tion at relatively low cost. The cost of acquiring ly will be used more frequently by those who need more detailed levels of land use and land cover data DEVELOPING THE CLASSIFICATION SYSTEM 7 might prohibit including such data on large-scale they are useful in categorizing data obtained from maps over extensive areas. remote sensor sources. Recent experiments (Stevens and others, 197 4) The definition of Urban or Built-up Land, for ex­ with Levels I and II land use data referenced to ample, includes those uses similarly classified (Woo­ 1:24,000 topographic maps have been conducted by ten and Anderson, 1957) by the U.S. Department of researchers of the Maps and Surveys Branch of the Agriculture, plus the built-up portions of major Tennessee Valley Authority in conjunction with the recreational sites, public installations, and other Marshall Space Flight Center and Oak Ridge Na­ similar facilities. Agricultural land has been defined tional Laboratories. Quite satisfactory results have to include Cropland and Pasture; Orchards, Groves, been obtained when interpreting land use from high­ Vineyards, Nurseries, and Ornamental Horticultural altitude photography. In areas of considerable ter­ Areas; and Confined Feeding Operations as the prin­ rain relief a stereoplotter was used to avoid scale cipal components. Certain land uses such as pasture, problems. however, cannot be separated consistently and ac­ The categories proposed at Level II cannot all be curately by using the remote sensor data sources interpreted with equal reliability. In parts of the appropriate to the more generalized levels of the United States, some categories may be extremely classification. The totality of the category thus close­ difficult to interpret from high-altitude aircraft ly parallels the U.S. Department of Agriculture defi­ imagery alone. Conventional aerial photography and nition of agricultural land. sources of information other than remote sensor The primary definition of Forest Land employed data may be needed for interpretation of especially for use with data acquired by remote sensors ap­ complex areas. On the basis of research and testing proximates that used by the U.S. Forest Service ( un­ carried out in the U.S.G.S. Geography Program's published manual), with the exception of those Central Atlantic Regional Ecological Test Site brush and shrub-form types such as chaparral and mesquite, which are classed as forest land by the (CARETS) Project, the Phoenix Pilot Project, and in land use mapping for the Ozarks Regional Com­ Forest Service because of their importance in water­ shed control. Because of their spectral response,, mission (U.S. Geological Survey, 1973), it has been these generally are grouped with Rangeland types in determined that the cost of using such supplemen­ tary information can be held to reasonable levels. classifications of vegetation interpretable from re­ mote sensing imagery. At Level III, which is beyond the scope of the The principal concept by which certain types of present discussion, use of substantial amounts of cover are included in the Rangeland category, and supplemental information in addition to some re­ which separates rangeland from pasture land, is motely sensed information at scales of 1 : 15,000 to that rangeland has a natural climax plant cover of 1:40,000 should be anticipated. Surprisingly de­ native grasses, forbs, and shrubs which is potential­ tailed inventories may be undertaken, and by using ly useful as a grazing or forage resource (U.S. Con­ both remotely sensed and supplemental information, gress, 1936; U.S. Department of Agriculture, 1962, most land use and land cover types, except those of 1971). Although these rangelands usually are not very complex urban areas or of thoroughly hetero­ seeded, fertilized, drained, irrigated, or cultivated, geneous mi:ktures can be adequately located, meas­ if the forage cover is improved, it is managed pri­ ured, and coded. marily like native vegetation, and the forage re­ Level IV would call for much more supplemental source is regulated by varying the intensity and information and remotely sensed data at a much seasonality of grazing (Stoddard and Smith, 1955). larger scale. Since the typical cropland practices mentioned just above are characteristics of some pasture lands, these DEVELOPING THE CLASSIFICATION SYSTEM pasture lands are similar in image signature to crop­ land types. In developing the classification system, every ef­ The definition of Wetland incorporates the major fort has been made to provide as much compatibility elements of the original U.S. Department of the as possible with other classification systems current­ Interior definition (Shaw and Fredine, 1956) as well ly being used by the various Federal agencies in­ as the combined efforts of the U.S.G.S. working volved in land use inventory and mapping. Special group on wetlands definition. attention has been paid to the definitions of land use Table 1 presents a general summary of land use categories used by other agencies, to the extent that compiled every 5 years by the Economic Research 8 A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA

Service of the U.S. Department of Agriculture and TABLE 2.-Land use and land cover classification system for supplemented from other sources. These statistics, use with remote sensor data which are available only for States, are provided by Level I Level II 1 Urban or Built-up Land 11 Residential. the various government agencies which compile in­ 12 Commercial and Services. formation on some categories of land use, several of 13 Industrial. 14 Transportation, Communi­ which parallel the U.S.G.S. land use classification cations, and Utilities. system. 15 Industrial and Commercial Complexes. TABLE 1.-Major uses of land, United States, 1969 1 16 Mixed Urban or Built-up Land. Acres Hectares 17 Other Urban or Built-up (mil­ (mil­ Per­ Land. lions) lions) cent 2 Agricultural Land 21 Cropland and Pasture. Cropland 472 191 20.9 22 Orchards, Groves, Vine­ yards, Nurseries, and Cropland used for crops ___ _ 333 135 Ornamental Horticultural Cropland harvested ______286 116 Areas. Crop failure ------­ 6 2 23 Confined Feeding Opera- Cultivated summer fallow_ 41 17 tions. Soil improvement crops and 24 Other Agricultural Land. idle cropland ------­ 51 21 3 Rangeland 31 Herbaceous Rangeland. Cropland used only for pasture 88 35 32 Shrub and Brush Range- Grassland pasture and range 2 ___ _ 604 245 26.7 land. 33 Mixed Rangeland. Forest land ------723 293 31.9 4 Forest Land 41 Deciduous Forest Land. Grazed ------198 80 42 Evergreen Forest Land. 525 213 Not grazed3 ------43 Mixed Forest Land. Special uses ------178 72 7.9 5 Water 51 Streams and Canals. Urban areas ------35 14 Transportation areas ______52 Lakes. 26 11 53 Reservoirs. Rural parks ------­ 49 19 54 Bays and Estuaries. Wildlife refuges ------­ 32 13 National defense, flood control, 6 Wetland 61 Forested Wetland.. and industrial areas ------­ 26 11 62 N onforested Wetland. State-owned institutions and 7 Barren Land 71 Dry Salt Flats. misc.ellaneous other uses __ _ 2 1 72 Beaches. Farmsteads, farm roads, 73 Sandy Areas other than and lanes ------­ 8 3 Beaches. Miscellaneous land"' ------287 116 12.6 74 Bare Exposed Rock. 75 Strip MinPs. Quarries, and 1 Frey, H. T., 1973. Does not include area covered by water in streams Gravel Pits. more than lfs of a mile in width and lakes, reservoirs, and so forth of more than 40 acres in size. 76 Transitional Areas. 2 Includes pasture that is to be included with cropland in the U.S.G.S. 77 Mixed Barren Land. classification system. 3 Except for urban and built-up areas and transportation uses, these 8 Tundra 81 Shrub and Brush Tundra. special uses will be classified by dominant cover under the U.S.G.S. classi­ 82 HerbBC'eous Tundra. fication system. 83 Bare Ground Tundra. 4 Tundra, glaciers, and icefields, marshes, open swamps, bare rock areas, deserts, beaches, and other miscellaneous land. 84 "-.. et Tundra. 85 Mixed Tundra. The land use and land cover classification system 9 Perennial Snow or Ice 91 Perennial Snowfields. 92 Glaciers. presented in this report (table 2) includes only the more generalized first and second levels. The system studies will depend on the improvement that should satisfies the three major attributes of the classifica­ result from widespread use of the system. tion process as outlined by Grigg ( 1965) : ( 1) it As further advances in technology are made, it gives names to categories by simply using accepted may be necessary to modify the classification system terminology; (2) it enables information to be trans­ for use with automatic data analysis. The LANDSAT mitted; and (3) it allows inductive generalizations and Skylab missions and the high-altitude aircraft to be made. The classification system is capable of program of the National Aeronautics and Space Ad­ further refi11ement on the basis of more extended and ministration have offered opportunities for nation­ varied use. At the more generalized levels it should wide testing of the feasibility of using this classifica­ meet the principal objective of providing a land use tion system to obtain land use information on a and land cover classification system for use in land uniform basis. u_se planning and management activities. Attainment The approach to land use and land cover classifi­ of the more fundamental and long-range objective cation embodied in the system described herein is of providing a standardized system of land use and "resource oriented," in contrast, for example, with land cover classification for national and regional the "people orientation" of the "Standard Land Use USING THE CLASSIFICATION SYSTEM 9

Coding Manual," developed by the U.S. Urban Re­ generated which identifies ownership-management newal Administration and the Bureau of Public units such as farms or ranches or relating detached Roads ( 1965). For the most part the Manual is uses, included in a specific ownership complex, to the derived from the "Standard Industrial Classification parent activity. For example, warehouses cannot be Code" established and published by the former Bu­ related to retail sales when the two occurrences are reau of the Budget (U.S. Executive Office of the separated spatially. The actual cover and related uses President, 1957). are mapped in each case, rather than injecting The people-oriented system of the '(Standard Land inference into the inventory process. Use Coding Manual" assigns seven of the nine gen­ Inferences used for prediction could cause prob­ eralized first level categories to urban, transporta­ lems for the land use interpreter where land use is tion, recreational, and related uses of land, which clearly in transition, with neither the former use nor account for less than 5 percent of the total area of the future use actually being present. In most such the United States (tables 1 and 3). Although there cases, it is tempting to speculate on future use,, but is an obvious need for an urban-oriented land use all that can actually be determined in such wide­ classification system, there is also a need for a ranging situations is that change is occurring. Large resource-oriented classification system whose pri­ clear-cut areas in the southeastern forests, for ex­ mary emphasis would be the remaining 95 percent of ample, are not always returned to forests and might the United States land area. The U.S.G.S. classifica­ assume any of a variety of future uses, such as a tion system described in this report addresses that residential subdivision, an industrial site, an area of need, with eight of the nine Level I categories treat­ cropland, or a phosphate mine. The "sagebrush sub­ ing land area of the United States that is not in division" of the Southwest may have all the potential urban or built-up areas. Six of the first level cate­ earmarks of future settlement, such as carefully gories in the standard land use code are retained platted streets, and yet never experience any con­ under Urban or Built-up at Level II in the U.S.G.S. struction. Such cleared open areas should be identi­ system. Even though the standard land use code and fied as "Transitional Areas." the U.S.G.S. classification differ considerably in their Since Level II will probably be most appropriate major emphases, a marked degree of compatibility for statewide and interstate regional land use and between these two systems exists at the more gen­ land cover compilation and mapping, and since Level eralized levels and even at the more detailed levels. II categories can be created by aggregating similar Level III categories, the Level II categorization may TABLE 3.-Standard land use code-first level categm·ies 1 be considered to be the fulcrum of the classification 1. Residential. system. The classification system may be entered at 2. Manufacturing (9 second level categories included). 3. Manufacturing ( 6 second level categories included). the particular level appropriate to the individual 4. Transportation, communications, and utilities. user, and the information generated may be added 5. Trade. 6. Services. together with data generated by others to form an 7. Cultural, entertainment, and recreation. aggregate category at the next higher level. As an 8. Resource production and extraction. 9. Undeveloped l.and and water areas. example, if a local planning group had devised a 1 Standard land use coding manual, 1965, p. 29. Level III classification of a particular group of land uses and had included sufficient definitional informa­ USING THE CLASSIFICATION SYSTEM tion of their land use categories, their data could be The use of the same or similar terminology does compiled into a larger inventory by a state or re­ not automatically guarantee that the land use data gional planning group compiling data by use of the collected and coded according to two systems will be Level II categories. Such data,-in turn, could serve as entirely compatible. The principal points of depar­ part of the data base for a national inventory. ture between other classifications and the U.S.G.S. Seldom is it necessary to inventory land uses at system originate because of the emphasis placed on the more detailed levels, even for local planning. remote sensing as the primary data source used in Having greater detail does, however, provide flexi­ the U.S.G.S. classification system. Beca.use of this bility in manipulating the data when several differ­ emphasis, activity must be interpreted using land ent purposes must be served. The cost of interpret­ cover as the principal surrogate, in addition to the ing, coding, and recording land use data at the more image interpreter's customary references to pattern, detailed levels is necessarily greater than if the data geographic location, and so forth. This process neces­ were handled at more generalized levels. This extra sarily precludes the possibility of information being cost reflects the increase in cost of remote sensor and 10 A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA collateral data acquired at larger scales, as well as Stewart (1965), "One man's miscellany is another the increase in interpretation costs. man's prime concern." No one would consider pub­ The U.S.G.S. classification system provides flexi­ lishing a map of current land use of any part of the bility in developing categorization at the more de­ Western United States without having irrigated tailed levels. Therefore, it is appropriate to illustrate land as a major category. With the flexibility inher­ the additive properties of the system and to provide ent in this classification system, an accommodation examples for users wishing to develop more detailed of this type of need can be made easily, provided categorization. The several examples given below that irrigated land is mapped or tabulated as a dis­ represent possible categorizations. Users should not crete unit which can be aggregated into the more consider themselves limited to categories such as general categories included in the framework of the these but should develop categories of utmost utility classification. A possible restructuring which would to their particular needs. It should be emphasized accommodate the desire to present irrigated land as that, whatever categories are used at the various a major category would be: classification levels, special attention should be given Irrigated agricultural land Nonirrigated ag.ricultural land to providing the potential users of the data with suffi­ Cropland Cropland cient information so that they may either compile Pasture Pasture the data into more generalized levels or aggregate Orchards, Groves and so forth Orchards, Groves and so forth more detailed data into the existing classes. DEFINITIONS One example of subcategorization of Residential Land as keyed to the standard land use code would An attempt has been made to include sufficient be: detail in the de·finitions presented here to provide a

Level I Level II Level III general understanding of what is included in each L. Urban or 11. Residential. 111. Single-family Units. category at Levels I and II. Many of the uses de­ Built-up 112. Multi-family Units. scribed in aetail will not be detectable on small-scale 113. Group Quarters. 114. Residential Hotels. aerial photographs. However, the detail will aid in 115. Mobile Home Parks. the interpretation process, and the additional infor­ 116. Transient Lodgings. 117. Other. mation will be useful to those who have large-scale aerial photographs and other supplemental informa­ This particular breakdown of "Residential" em­ ploys criteria of capacity, type, and permanency of tion available. residence as the discriminating factors among 1. URBAN OR BUILT-UP LAND classes. Criteria applied to other situations could pos­ sibly include density of dwellings, tenancy, age of Urban or Built-up Land is comprised of areas of construction, and so forth. Obviously, such a Level intensive use with much of the land covered by struc­ III categorization would require use of supplemental tures. Included in this category are cities, towns, vil­ information. Users desiring Level IV information lages, strip developments along highways, transpor­ could employ a variety of additional criteria in dis­ tation, power, and communications facilities, and criminating among land uses,but it can be seen that areas such as those occupied by mills, shopping cen­ the element which allows aggregation and transfer ters, industrial and commercial complexes, and insti­ between categories is the proper description of what tutions that may, in some instances, be isolated from is included in each individual category at whatever urban areas. level the data are being classified. As development progresses, land having less inten­ The Level II category, Cropland and Pasture, may sive or nonconforming use may be located in the be simply subdivided at Level III. midst of Urban or Built-up areas and will generally Level II Level III be included in this category. Agricultural land, for­ 21. Cropland and Pasture. 211. Cropland. est, wetland, or water areas on the fringe of Urban 212. Pastul"€1. or Built-up areas will not be included except where Some users may wish such additional criteria em­ they are surrounded and dominated by urban devel­ ployed at Level III as degree of activity or idleness opment. The Urban or Built-up category takes prece­ or degree of improvement, while others may place dence over others when the criteria for more than such items in Levels IV or V. What may be a primary one category are met. For example, residential areas category for one user group may be of secondary that have sufficient tre·e cover to meet Forest Land importance to another. As stated by Clawson and criteria will be placed in the Residential category. DEFINITIONS 11

11. RESIDENTIAL to cities; junkyards; resorts; and so forth. The main buildings, secondary structures, and areas support­ Residential land uses range from high density, ing the basic use are all included-office buildings, represented by the multiple-unit structures of urban warehouses, driveways, sheds, parking lots, land­ cores, to low density, where houses are on lots of scaped areas, and waste disposal areas. more than an acre, on the periphery of urban expan­ Commercial areas may include some noncommer­ sion. Linear residential developmenh; along trans­ cial uses too small to be separated out. Central busi­ portation routes extending outwar::l from urban ness districts commonly include some institutions, areas should be included as residential appendages such as churches and schools, and commercial strip to urban centers, but care must be taken to dis­ developments may include some residential units. tinguish them from commercial strips in the same When these noncommercial uses exceed one-third-of locality. The residential strips generally have a uni­ the total commercial area, the Mixed Urban or Built­ form size and spacing of structures, linear drive­ up category should be used. There is no separate ways, and lawn areas; the commercial strips are ca~tegory for recreational land uses at Level II since more likely to have buildings of different sizes and most recreational activity is pervasive throughout spacing, large driveways, and parking areas. Resi­ many other land uses. Selected areas are predomi­ dential development along shorelines is also linear nantly recreation oriented, and some of the more dis­ and sometimes extends back only one residential tinctive occurrences such as drive-in theaters can be parcel from the shoreline to the first road. identified on remote sensor imagery. Most recrea­ Areas of sparse residential land use, such as farm­ tional activity, however, necessarily will be identified steads, will be included in categories to which they using supplemental information. Recreational facili­ are related unless an appropriate compilation scale ties that form an integral part of an institution is being used to indicate such uses separately. Rural should be included in this category. There is usually residential and recreational subdivisions, however, a major visible difference in the form of parking are included in this category, since the land is almost facilities, arrangements for traffic flow, and the gen­ totally committed to residential use, even though it eral association of buildings and facilities. The in­ may have forest or range types of cover. In some tensively developed sections of recreational areas places, the boundary will be clear where new housing would be included in the Commercial and Services developments abut against intensively used agricul­ category, but extensive parts of golf courses, riding tural areas, but the boundary may be vague and diffi­ areas, ski areas, and so forth would be included in cult to discern when residential development occurs the Other Urban or Built-up category. in small isolated units over an area of mixed or less Institutional land uses, such as the various educa­ intensive uses. A careful evaluation of density and tional, religious, health, correctional, and military the overall relation of the area to the total urban facilities are also components of this category. All complex must be made. buildings, grounds, and parking lots that compose Re~idential sections which are integral parts of the facility are included within the institutional unit, other uses may be difficult to identify. Housing situa­ but areas not specifically related to the purpose of tions such as those existing on military bases, at col­ the institution should be placed in the appropriate leges and universities, living quarters for laborers category. Auxiliary land uses, particularly residen­ near a work base, or lodging for employees of agri­ tial, commercial and services, and other supporting cultural field operations or resorts thus would be land uses on a military base would be included in this placed within the Industrial, Agricultural, or Com­ category, but agricultural areas not specifically as­ mercial and Services categories. sociated with correctional, educational, or religious institutions are placed in the appropriate agricul­ 12. COMMERCIAL AND SERVICES tural category. Small institutional units, as, for ex­ Commercial areas are those used predominantly ample, many churches and some secondary and ele­ for the sale of products and services. They are often mentary schools, would be mappable only at large abutted by residential, agricultural, or other con­ scales and will usually be included within another trasting uses which help define them. Components of category, such as Residential. the Commercial and Services category are urban cen­ tral business districts ; shopping centers, usually in 13. INDUSTRIAL suburban and outlying areas; commercial strip de­ Industrial areas include a wide array of land uses velopments along major highways and access routes from light manufacturing to heavy manufacturing 12 A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA

plants. Identification of light industries-those fo­ Highways and railways are characterized by areas cused on design, assembly, finishing, processing, and of activity connected in linear patterns. The high­ packaging of products-can often be based on the ways include rights-of-way, areas used for inter­ type of building, parking, and shipping arrange­ changes, and service and terminal facilities. Rail ments. Light industrial areas may be, but are not facilities include stations, parking lots, roundhouses, necessarily, directly in contact with urban areas; repair and switching yards, and related areas, as many are now found at airports or in relatively open well as overland track and spur connections of suffi­ country. Heavy industries use raw materials such cient width for delineation at mapping scale. as iron ore, timber, or coal. Included are steel mills, Airports, seaports, and major lakeports are iso­ pulp and lumber mills, electric-power generating lated areas of high utilization, usually with no well­ stations, oil refineries and tank farms, chemical defined intervening connections, although some ports plants, and brickmaking plants. Stockpiles of raw are connected by canals. Airport facilities include materials and waste-product disposal areas are usu­ the runways, intervening land, terminals, service ally visible, along with transportation facilities buildings, navigation aids, fuel storage, parking lots) capable of handling heavy materials. and a limited buffer zone. Terminal facilities general­ Surface structures associated with mining opera­ ly include the associated freight and warehousing tions are included in this category. Surface struc­ functions. Small airports (except those on rotated tures and equipment may range from a minimum of farmland), heliports, and land associated with sea­ a loading device and trucks to extended areas with plane bases may be identified if mapping scale per­ access roads, processing facilities, stockpiles, storage mits. Port areas include the docks, shipyards, dry­ sheds, and numerous vehicles. Spoil material and slag docks, locks, and waterway control structures. heaps usually are found within a short trucking dis­ Communications and utilities areas such as those tance of the major mine areas and may be the key involved in processing, treatment, and transporta­ indicator of underground mining operations. Uni­ tion of water, gas, oil, and electricity and areas used form identification of all these diverse extractive for airwave communications are also included in this uses is extremely difficult from remote sensor data category. Pumping stations, electric substations, and alone. Areas of future reserves are included in the areas used for radio, radar, or television antennas appropriate present-use category, such as Agricul­ are the major types. Small facilities, or those associ­ tural Land or Forest Land, regardless of the ex­ ated with an industrial or commercial land use, are pected future use. included within the larger category with which they are associated. Long-distance gas, oil, electric, tele­ 14. TRANSPORTATION, COMMUNICATIONS, phone, water, or other transmission facilities rarely AND UTILITIES constitute the dominant use of the lands with which they are associated. The land uses included in the Transportation, Communications, and Utilities category occur to 15. INDUSTRIAL AND COMMERCIAL COMPLEXES some degree within all of the other Urban or Built­ The Industrial and Commercial Complexes cate­ up categories and actually can be found within many gory includes those industrial and commercial land other categories. Unless they can be mapped sepa­ uses that typically occur together or in close func­ rately at whatever scale is being employed, they tional proximity. Such areas commonly are labeled usually are considered an integral part of the land with terminology such as "Industrial Park," but use within which they occur. For that reason, any since functions such as warehousing, wholesaling, statistical summary of the area of land uses in this and occasionally retailing may be found in the same category typically represents only a partial data set. structures or nearby, the more inclusive category Statistical area summaries of such land uses aggre­ title has been adopted. gated from Levels III and IV, though, would include Industrial and Commercial complexes have a defi­ more accurate area estimates. nite remote sensor image signature which allows Major transportation routes and areas greatly their separation from other Urban or Built-up land influence other land uses, and many land use bound­ uses. Because of their intentional development as dis­ aries are outlined by them. The types and extent of crete units of land use, they may border on a wide transportation facilities in a locality determine the variety of other land use types, from Residential degree of access and affect both the present and po­ Land to Agricultural Land to Forest Land. If the tential use of the area. separate functions included in the category are iden- DEFINITIONS 13 tified at Levels III or IV using supplemental data or However, pasture and other lands where such equip­ with ground survey, the land use researcher has the ment is used infrequently may not show as well­ discretion of aggregating these functions into the defined shapes as other areas. These distinctive geo­ appropriate Level II Urban or Built-up categories or metric patterns are also characteristic of Urban or retaining the unit as an Industrial and Commercial Built-up Lands because of street layout and develop­ Complex. ment by blocks. Distinguishing between Agricultural and Urban or Built-up Lands ordinarily should be 16. MIXED URBAN OR BUILT-UP LAND possible on the basis of urban-activity indicators and The Mixed Urban or Built-up category is used for the associated concentration of population. The num­ a mixture of Level II Urban or Built-up uses where ber of building complexes is smaller and the density individual uses cannot be separated at mapping scale. of the road and highway network is much lower in Where more than one-third intermixture of another Agricultural Land than in Urban or Built-up Land. use or uses occurs in a specific area, it is classified Some urban land uses, such as parks and large ceme­ as Mixed Urban or Built-up Land. Where the inter­ teries, however, may be mistaken for Agricultural mixed land use or uses total less than one-third of Land, especially when they occur on the periphery of the specific area, the category appropriate to the the urban areas. dominant land use is applied. The interface of Agricultural Land with other This category typically includes developments categories of land use may sometimes be a transition along transportation routes and in cities, towns, and zone in which there is an intermixture of land uses built-up areas where separate land uses cannot be at first and second levels of categorization. Where mapped individually. Residential, Commercial, In­ farming activities are limited by wetness, the exact dustrial, and occasionally other land uses may be boundary also may be difficult to locate, and Agricul­ included. A mixture of industrial and commercial tural Land may grade into Wetland. When the pro­ uses in Industrial and Commercial Complexes as de­ duction of agricultural crops is not hindered by wet­ fined in category 15 are not included in this category. land conditions, such cropland should be included in Farmsteads intermixed with strip or cluster settle­ the Agricultural category. This latter stipulation ments will be included within the built-up land, but also includes those cases in which agricultural crop other agricultural land uses should be excluded. production depends on wetland conditions, such as the flooding of ricefields or the development of cran­ 17. OTHER URBAN OR BUlL T-UP LAND berry bogs. When lands produce economic commodi­ Other Urban or Built-up Land typically consists of ties as a function of their wild state such as wild uses such as golf driving ranges, zoos, urban parks, rice, cattails, or certain forest products ·commonly cemeteries, waste dumps, water-control structures associated with wetland, however, they should be in­ and spillways, the extensive parts of such uses as cluded in the Wetland category. Similarly, when wet­ golf courses and ski areas, and undeveloped land lands are drained for agricultural purposes, they within an urban setting. Open land may be in very should be included in the Agricultural Land cate­ intensive use but a use that does not require struc­ gory. When such drainage enterprises fall into dis­ tures, such as urban playgrounds, botanical gardens, use and if wetland vegetation is reestablished, the or arboreta. The use of descriptions such as "idle land reverts to the Wetland category. land," "vacant land," or '"open land" should be The Level II categories of Agricultural Land are: avoided in categorizing undeveloped lands within Cropland and Pasture; Orchards, Groves, Vineyards, urban areas on the basis of the use of remote sensor Nurseries, and Ornamental Horticultural Areas ; data, since information generally is not available to Confined Feeding Operations; and Other Agricul­ the interpreter to make such a refinement in tural Land. categorization. 21. CROPLAND AND PASTURE 2. AGRICULTURAL LAND The several components of Cropland and Pasture Agricultural Land may be defined broadly as land now used for agricultural statistics include: crop- used primarily for production of food and fiber. On ! land harvested, including bush fruits; cultivated high-altitude imagery, the chief indications of agri­ summer-fallow and idle cropland; land on which cultural activity will be distinctive geometric field crop failure occurs; cropland in soil-improvement and road patterns on the landscape and the traces grasses and legumes; cropland used only for pasture produced by livestock or mechanized equipment. in rotation with crops; and past~re on land more or 14 A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA

less permanently used for that purpose. From imag­ climatological factors needed for these operations: ery alone, it generally is not possible to make a dis­ water bodies in close proximity which moderate the tinction between Cropland and Pasture with a high effects of short duration temperature fluctuations.; degree of accuracy and uniformity, let alone a dis­ site selection for air drainage on sloping land; and tinction among the various components of Cropland deep well-drained soils on slopes moderate enough to (Hardy, Belcher, and Phillips, 1971). Moreover, permit use of machinery. Isolated small orchards, some of the components listed represent the condi­ such as the fruit trees on the family farm, usually tion of the land at the end of the growing season are not recognizable on high-altitude imagery and and will not apply exactly to imagery taken at other are, therefore, not included. times of the year. They will, however, be a guide to identification of Cropland and Pasture. Brushland 23. CONFINED FEEDING OPERATIONS in the Eastern States, typically used to some extent Confined Feeding Operations are large, specialized for pasturing cattle, is included in the Shrub-Brush­ livestock production enterprises, chiefly beef cattle land Rangeland category since the grazing activity feedlots, dairy operations with confined feeding, and is usually not discernible on remote sensor imagery large poultry farms, but also including hog feedlots. appropriate to Levels I and II. This activity possibly These operations have large animal populations re­ might be distinguished on low-altitude imagery. Such stricted to relatively small areas. The result is a con­ grazing activities generally occur on land where crop centration of waste material that is an environmental production or intensive pasturing has ceased, for concern. The waste-disposal problems justify a sepa­ any of a variety of reasons, and which has grown rate category for these relatively small areas. Con­ up in brush. Such brushlands often are used for fined Feeding Operations have a built-up appear­ grazing, somewhat analogous to the extensive use of ance, chiefly composed of buildings, much fencing, rangelands in the West. access paths, and waste-disposal areas. Some are Certain factors vary throughout the United States, located near an urban area to take advantage of and this variability also must be recognized; field transportation facilities and proximity to process­ size depends on topography, soil types, sizes of ing plants. farms, kinds of crops and pastures, capital invest­ Excluded are shipping corrals and other tempo­ ment, labor availability, and other conditions. Irri­ rary holding facilities. Such occurrences as thor­ gated land in the Western States is recognized easily oughbred horse farms generally do not have the in contrast to Rangeland, but in the Eastern States, animal population densities which would place them irrigation by use of overhead sprinklers generally in this category. cannot be detected from imagery unless distinctive 24. OTHER AGRICULTURAL LAND circular patterns are created. Drainage or water con­ trol on land used for cropland and pasture also may Other land uses typically associated with the first create a recognizable pattern that may aid in identi­ three categories of Agricultural Land are the princi­ fication of the land use. In areas of quick-growing pal components of the Other Agricultural Land cate­ crops, a field may appear to be in nonagricultural use gory. They include farmsteads, holding areas for unless the temporary nature of the inactivity is livestock such as corrals, breeding and training fa... recognized. cilities on horse farms, farm lanes and roads, ditches and canals, small farm ponds, and similar uses. Such 22. ORCHARDS, GROVES, VINEYARDS, NURSERIES, occurrences generally are quite small in area and AND ORNAMENTAL HORTICULTURAL AREAS often uninterpretable by use of high-altitude data. Orchards, groves, and vineyards produce the vari­ Even when they are interpretable from such data, it ous fruit and nut crops. Nurseries and horticultural may not be feasible to map them at smaller presenta­ areas, which include floricultural and seed-and-sod tion scales, which generally results in their inclusion areas and some greenhouses, are used perennially for with adjacent agricultural use areas. This category those purposes. Tree nurseries which provide seed­ should also be used for aggregating data for land lings for plantation forestry also are included here. uses derived at more detailed levels of classification. Many of these areas may be included in another cate­ gory, generally Cropland and Pasture, when identifi­ 3. RANGELAND cation is made by use of small-scale imagery alone. Rangeland historically has been defined as land Identification may be aided by recognition of the where the potential natural vegetation is predomi­ combination of soil qualities, topography, and local nantly grasses, grasslike plants, forbs, or shrubs and DEFINITIONS 15 where natural herbivory was an important influence palms (Sabal palmetto), and shrubs (Shelford, in its precivilization state. Management techniques 1963). Those palmetto prairie areas now in im­ which associate soil, water, and forage-vegetation proved pasture would not be included in this cate­ resources are more .suitable for rangeland manage­ gory, nor would the herbaceous varieties of tundra ment than are practices generally used in managing vegetation. pastureland. Some rangelands have been or may be 32. SHRUB AND BRUSH RANGELAND seeded to introduced or domesticated plant species. Most of the rangelands in the United States are in The typical shrub occurrences are found in those the western range, the area to the west of an irregu­ arid and semiarid regions characterized by such lar north-south line that cuts through the Dakotas, xerophytic vegetative types with woody stems as big Nebraska, Kansas, Oklahoma, and Texas. Range­ sagebrush (Artemisia tridentata), shadscale (Atri­ lands also are found in certain places historically not plex conje1·tijolia), greasewood (Sarcobatus vermi­ included in the western range, such as the Flint culatus), or creosote bush (Larrea divaricata) and Hills, the Southeastern States, and Alaska. The his­ also by the typical desert succulent xerophytes, such torical connotation of Rangeland is expanded in this as the various forms of Cactus (Kuchler, 1964). classification to include those areas in the Eastern When bottom lands and moist flats are characterized States which commonly are called brushlands. by dense stands of typical wetland species such as The Level II categories of Rangeland are : Herba­ mesquite (Prosopis), they are considered Wetland. ceous Range, Shrub and Brush Rangeland, and Where highly alkaline soils are present, halophytes Mixed Rangeland. such as desert saltbush (A triplex) may occur. The type, density, and association of these various species 31. HERBACEOUS RANGELAND are useful as indicators of the local hydrologic and The Herbaceous Rangeland category encompasses pedologic environments. Also included in this cate­ lands dominated by naturally occurring grasses and gory are chaparral, a dense mixture of broadleaf forbs as well as those areas of actual rangeland evergreen schlerophyll shrubs, and the occurrences which have been modified to include grasses and of mountain mahogany (Cercocarpus ledifolius) and forbs as their principal cover, when the land is man­ scrub oaks (Quercus). aged for rangeland purposes and not managed using The eastern brushlands are typically former crop­ practices typical of pastureland. It includes the tall lands or pasture lands (cleared from original forest grass (or true prairie) , short grass, bunch grass or land) which now have grown up in brush in transi­ palouse grass, and desert grass regions. Respective­ tion back to forest land to the extent that they are no ly, these grass regions represent a sequence of de­ longer identifiable as cropland or pasture from re­ clining amounts of available moisture. Most of the mote sensor imagery. Many of these brushlands are tall grass region has been plowed for agriculture and grazed in an extensive manner by livestock and pro­ the remaining tall grass range is now in North Da­ vide wildlife habitat. These areas usually ~main as kota, Nebraska, southern Kansas and Oklahoma, and part of the farm enterprise, even though not being the Texas Coastal Plain. Short grass rangeland oc­ used at their former levels of intensity. Eastern curs in a strip about 300 miles (500 km) wide from brushland areas traditionally have not been included the Texas Panhandle northward to the Dakotas in the rangeland concept because of their original where it widens to cover the western half of the forested state prior to clearing for cropland or pas­ Dakotas, the eastern three-fourths of Montana, and ture and generally have been summarized statistical­ the eastern third of Wyoming. Bunch grass and ly with pastureland. Because they function now pri­ desert grass are found in many locations, represent­ marily as extensive grazing land, they are included ing transitional situations to desert shrub. Typical here as part of the Rangeland category. After suffi­ occurrences of grasslands include such species as the cient forest growth has occurred, they should be various bluestems (Andropogon), grama grasses classified as either Deciduous, Evergreen, or Mixed (Bouteloua), wheatgra.sses (Agropyron), needle- Forest Land. Those occurrences of shrubs and brush grasses (Stipa), and fescues (Festuca). which are part of the Tundra are not included under Rangeland. This category also includes the palmetto prmne areas of south-central Florida, which consist mainly 33. MIXED RANGELAND of dense stands of medium length and tall grasses When more than one-third intermixture of e~ither such as wiregrass (Aristida stricta) and saw pal­ herbaceous or shrub and brush rangeland species oc­ mettos (Seronoa t·ipens), interspersed occasional curs in a specific area, it is classified as Mixed 16 A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA

Rangeland. Where the intermixed land use or uses differentiate these three categories effectively, se­ total less than one-third of the specific area, the quential data, or at least data acquired during the category appropriate to the dominant type of Range­ period when deciduous trees are bare, generally will land is applied. Mixtures of herbaceous and shrub or be necessary. brush tundra plants are not considered Rangeland. 41. DECIDUOUS FOREST LAND 4. FOREST LAND Deciduous Forest Land includes all forested areas Forest Lands have a tree-crown areal density having a predominance of trees that lose their leaves (crown closure percentage) of 10 percent or more, at the end of the frost-free season or at the begin­ are stocked with trees capable of producing timber ning of a dry season. In most parts of the United or other wood products, and exert an influence on States, these would be the hardwoods such as oak the climate or water regime. Forest Land generally (Querc·us), maple (Acer), or hickory (Carya) and can be identified rather easily on high-altitude imag­ the "soft" hardwoods, such as aspen (Populus tremu­ ery, although the boundary between it and other loides) ( Shelford, 1963) . Tropical hardwoods are categories of land may be difficult to delineate included in the Evergreen Forest Land category. precisely. Deciduous forest types characteristic of Wetland, Lands from which trees have been removed to such as tupelo (Nyssa) or cottonwood (Populus less than 10 percent crown closure but which have deltoides), also are not included in this category. not been developed for other uses also are included. 42. EVERGREEN FOREST LAND For example, lands on which there are rotation cy­ cles of clearcutting and blockplanting are part of Evergreen Forest Land includes all forested areas Forest Land. On such lands, when trees reach mar­ in which the trees are predominantly those which ketable size, which for pulpwood in the Southeastern remain green throughout the year. Both coniferous United States may occur in 2 to 3 decades, there will and broad-leaved evergreens are included in this be large areas that have little or no visible forest category. In most areas, the coniferous evergreens growth. The pattern can sometimes be identified by predominate, but some of the forests of Hawaii are the presence of cutting operations in the midst of a notable exceptions. The coniferous evergreens are large expanse of forest. Unless there is evidence of commonly referred to or classified as softwoods. other use, such areas of little or no forest growth They include such eastern species as the longleaf should be included in the Forest Land category. pine (Pinus palustris), slash pine (Pinus ellioti), Forest land which is grazed extensively, as in the shortleaf pine (Pinus echinata), loblolly pine (Pinus Southeastern States, would be included in this cate­ taeda), and other southern yellow pines; various gory because the dominant cover is forest and the spruces (Picea) and balsam fir (Abies balsamea) ; dominant activities are forest related. Such activities white pine (PirntS strobus), red pine (Pinus resino­ could iorm the basis for Levels III or IV categoriza.. sa), and jack pine (Pinus banksiana); and hemlock tion. Lands that meet the requirements for Forest (Tsuga canadensis); and such western species as Land and also for an Urban or Built-up category Douglas-fir (Pseudotsttga menziesii), redwood (Se­ should be placed in the latter category. The only quoia sempe1·virens), ponderosa pine (Pinus monti­ exceptions in classifying Forest Land are those areas cola), Sitka spruce (Picea sitchensis), Engelmann which would otherwise be classified as Wetland if not spruce (Picea engelmanni), western redcedar (Thu­ for the forest cover. Since the wet condition is of ja plicata), and western hemlock (Tsuga heterophyl­ much interest to land managers and planning groups la) (Shelford, 1963). Evergreen species commonly and is so important as an environmental surrogate associated with Wetland, such as tamarack (Larix and control1 such lands are classified as Forested laricina) or black spruce (Picea mariana), are not Wetland. included in this category (Kuchler, 1964). Auxiliary concepts associated with Forest Land, 43. MIXED FOREST LAND such as wilderness reservation, water conservation, or ownership classification, are not detectable using Mixed Forest Land includes all forested areas remote sensor data. Such concepts may be used for where both evergreen and deciduous trees are grow­ creating categories at the more detailed levels when ing and neither predominates. When more than one­ supplemental information is available. third intermixture of either evergreen or deciduous At Level II, Forest Land is divided into three species occurs in a specific area, it is classified as categories: Deciduous, Evergreen, and Mixed. To Mixed Forest Land. Where the intermixed land use DEFINITIONS 17 or uses total less than one-third of the specified area, and so forth. Dams, levees, other water-control the category appropriate to the dominant type of structures, or the excavation itself usually will be Forest Land is applied, whether Deciduous or Ever­ evident to aid in the identification, although the green. water-control structures themselves and spillways are included in the Other Urban or Built-up Land 5. WATER category. The delineation- of water areas depends on the In most cases, reservoirs serve multiple purposes scale of data presentation and the scale and resolu­ and may include all of the land use functions just tion characteristics of the remote sensor data used mentioned. In certain cases like the Tennessee River, for interpretation of land use and land cover. (Water the entire length of the trunk stream is impounded. as defined by the Bureau of the Census includes all In such a situation, the stream exists as a stairstep areas within the land mass of the United States that series of impoundments with waterway, flood-con­ persistently are water covered, provided that, if trol, recreation, and power-generation functions but linear, they are at least lj8 mile (200 m) wide and, is still considered a reservoir, since the additional if extended, cover at least 40 acres (16 hectares) .) functions are the result of impoundment. For many purposes, agencies need information on the size and number of water bodies smaller than 54. BA¥S AND ESTUARIES Bureau of the Census minimums. These frequently Bays and Estuaries are inlets or arms of the sea can be obtained from small-scale remote sensor data that extend inland. They are included in this system with considerable accuracy. only when they are considered to be inland water and therefore are included within the total area of 51. STREAMS AND CANALS the United States. Those bay and estuarine water The Streams and Canals category includes rivers, areas classified as "other than inland water" are not creeks, canals, and other linear water bodies. Where included within the total area of the United States. the water course is interrupted by a control struc­ These "other than inland water" areas are adjacent ture, the impounded area will be placed in the to certain States and fall under their jurisdiction. Reservoirs category. They occur in primary bodies of water such as the The boundary between streams and other bodies Atlantic Ocean coastal waters, Chesapeake Bay, of water is the straight line across the mouth of the Delaware Bay, Long Island Sound, Gulf of Mexico, stream up to 1 nautical mile (1.85 km). Beyond that Pacific Ocean coastal waters, Puget Sound, the limit, the classification of the water body changes Straits of Georgia and Juan de Euca, Gulf of Alaska, to the appropriate categQry, whether it be Lakes, Bering Sea, Arctic Ocean coastal waters, and the Reservoirs, or Bays and Estuaries. These latter cate­ Great Lakes (U.S. Bureau of the Census, 1970). gories are used only if the water body is considered Only those bays and estuaries classified as inland to be "inland water" and therefore included in the water are included in this category. No category is total area of the United, States. No category is ap­ applied to offshore waters beyond the limits of Bays plied to waters classified as "other than inland and Estuaries. water " or offshore marine waters beyond the mouths of rivers '(U.S. Bureau of the Census, 1970). 6. WETLAND Wetlands are those areas where the water table is 52. LAKES at, near, or above the land surface for a significant Lakes are nonflowing, naturally enclosed bodies part of most years. The hydrologic regime is such of water, including regulated natural lakes but ex­ that aquatic or hydrophytic vegetation usually is cluding reservoirs. Islands that are too small to established, although alluvial and tidal flats may be delineate should be included in the water area. The non vegetated. Wetlands frequently are associated delineation of a lake should be based on the areal with topographic lows, even in mountainous regions. extent of water at the time the remote sensor data Examples of wetlands include marshes, mudflats, are acquired. and swamps situated on the shallow margins of bays, lakes, ponds, streams, and manmade impoundments 53. RESERVOIRS such as reservoirs. They include wet meadows or Reservoirs are artificial impoundments of water perched bogs in high mountain valleys and season­ used for irrigation, flood control, municipal water ally wet or flooded basins, playas, or potholes with supplies, recreation, hydroelectric power generation, no surface-water outflow. Shallow water areas 18 A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA where aquatic vegetation is submerged are classed ally flooded bottomland hardwoods, mangrove as open water and are not included in the Wetland swamps, shrub swamps, and wooded swamps in­ category. eluding those around bogs. Because Forested Wet- Extensive parts of some river flood plains qualify lands can be detected and mapped by the use of as Wetlands, as do regularly flooded irrigation over- seasonal (winter/summer) imagery, and because flow areas. These do not include agricultural land delineation of Forested Wetlands is needed for many where seasonal wetness or short-term flooding may environmental planning activities, they are sepa­ provide an important component of the total annual rated from other categories of Forest Land. soil moisture necessary for crop production. Areas The following are examples of typical vegetation in which soil wetness or flooding is so short-lived found in Forested Wetland. Wooded swamps and that no typical wetlands vegetation is developed southern flood plains contain primarily cypress properly belong in other categories. ( Taxodium), tupelo (Nyssa), oak.s (Quercus), and Cultivated wetlands such as the flooded fields red maple (Acer rubrum). Mangroves (Avicennia associated with rice production and developed cran- and Rhizophora) are dominant in certain subtropi­ berry bogs are classified as Agricultural Land. Un- cal Forested Wetland areas. Central and northern cultivated wetlands from which wild rice, cattails, flood plains are dominated by cottonwoods (Pop­ or wood products, and so forth are harvested, or ulus), ash (Fraxinus), alder (Alnus), and willow wetlands grazed by livestock, are retained in the , (Salix). Flood plains of the Southwest may be domi­ Wetland category. nated by mesquite (Prosopis), saltcedar (Tamarix), Remote sensor data provide the primary .source of seepwillow (Baccharis), and arrowweed (Pluchea). land use and vegetative cover information for the Northern bogs typically contain tamarack or larch more generalized levels of this classification system. (Larix), black spruce (Picea mariana), and heath Vegetation types and detectable surface water or shrubs (Ericaceae). Shrub swamp vegetation in­ soil moisture interpreted from such data provide the eludes alder (Alnus), willow (Salix), and button­ most appropriate means of identifying wetlands and bush (Cephalanthus occidentalis). wetland boundaries. Inasmuch a.s vegetation re­ sponds to changes in moisture conditions, remote 62. NONFORESTED WETLAND sensor data acquired over a period of time will allow N onforested Wetlands are dominated by wetland the detection of fluctuations in wetland conditions. herbaceous vegetation or are nonvegetated. These Ground surveys of soil types or the duration of wetlands include tidal and nontidal fresh, brackish, flooding may provide supplemental information to and salt marshes and nonvegetated flats and also be employed at the more detailed levels of classifica­ freshwater meadows, wet prairies, and open bogs. tion. The following are examples of vegetation asso­ Wetland areas drained for any purpose belong to ciated with Nonforested Wetland. Narrow-leaved other land use and land cover categories such as emergents such as cordgrass (Spartina) and rush Agricultural Land, Rangeland, Forest Land, or (Juncus) are dominant in coastal salt marshes. Both Urban or Built-up Land. When the drainage is dis­ narrow-leaved emergents such as cattail (Typha), continued and such use ceases, classification may bulrush (Scirpus), sedges ( Carex), sawgrass revert to Wetland. Wetlands managed for wildlife ( Cladium) and other grasses (for example, Pani­ purposes may show .short-term changes in land use cum and Zizaniopsis miliacea), and broad-leaved as different management practices are used but are emergents such as waterlily (Nuphar, Nymphea), properly classified Wetland. pickerelweed (Po·ntederia), arrow arum (Peltan­ Two separate boundaries are important with re­ dra), arrowhead (Sagittaria), water hyacinth spect to wetland discrimination: the upper wetland (Eichhornia crassipes), and alligatorweed (Altern- boundary above which practically any category of anthera philoxe1·oides) are typical of brackish to land use or land cover may exist, and the boundary freshwater locations. Mosses (Sphagnum) and between wetland and open water beyond which the sedges ( C m·ex) grow in wet meadows and bogs. appropriate Water category should be employed. 7. BARREN LAND Forested Wetland and N onforested Wetland are the Level II categories of Wetland. Barren Land is land of limited ability to support life and in which less than one-third of the area has 61. FORESTED WETLAND vegetation or other cover. In general, it is an area Forested Wetlands are wetlands dominated by of thin soil, sand, or rocks. Vegetation, if present, woody vegetation. Forested Wetland includes season- is more widely spaced and scrubby than that in the DEFINITIONS 19

Shrub and Brush category of Rangeland. Unusual 74. BARE EXPOSED ROCK conditions, such as a heavy rainfall, occasionally The Bare Exposed Rock category includes areas result in growth of a short-lived, more luxuriant of bedrock exposure, des,ert pavement, scarps, talus, plant cover. Wet, nonvegetated barren lands are in­ slides, volcanic material, rock glaciers, and other cluded in the Nonforested Wetland category. accumulations of rock without vegetative cover, with Land may appear barren because of man's activ­ the exception of such rock exposures occurring in ities. When it may reasonably be inferred from the tundra regions. data source that the land will be returned to its former use, it is not included in the Barren cate­ 75. STRIP MINES, QUARRIES, AND GRAVEL PITS gory but classified on the basis of its site and situ­ Those extractive mining activities that have sig­ ation. Agricultural land, for example, may be tem­ nificant surface expression are included in this cate­ porarily without vegetative cover because of crop­ gory. Vegetative cover and overburden are removed ping season or tillage practices. Similarly, industrial to expose such deposits as coal, iron ore, limestone, land may have waste and tailing dumps, and areas and copper. Quarrying of building and decorative of intensively managed forest land may have clear­ stone and recovery of sand and gravel deposits also cut blocks evident. result in large open surface pits. Current mining When neither the former nor the future use can activity is not always distinguishable, and inactive, be discerned and the area is obviously in a. state of unreclaimed, and active strip mines, quarries, bor­ land use transition, it is considered to be Barren row pits, and gravel pits are included in this cate­ Land, in order to avoid inferential errors. gory until other cover or use has been established, Level II categories of Barren Land are: Dry Salt after which the land would be classified in accord­ Flats, Beaches, Sandy Areas other than Beaches; ance with the resulting use or cover. Unused pits or Bare Exposed Rock; Strip Mines, Quarries, and quarries that have been flooded, however, are placed Gravel Pits; Transitional Areas; and Mixed Barren in the appropriate Water category. Land. 76. TRANSITIONAL AREAS 71. DRY SALT FLATS The Transitional Areas category is intended for Dry Salt Flats occurring on the flat-floored bot­ those areas which are in transition from one land toms of interior desert basins which do not qualify use activity to another. They are characterized by as Wetland are included in this category. On aerial the lack of any remote sensor information which photographs, Dry Salt Flats tend to appear white or would enable the land use interpreter to predict re­ light toned because of the high concentrations of liably the future use or discern the past use. All that salts at the surface as water has been evaporated, actually can be determined in these situations is resulting in a higher albedo than other adjacent that a transition is in progress, and inference about desert features. past or future use should be avoided. This transi­ tional phase occurs when, for example, forest lands 72. BEACHES are cleared for agriculture, wetlands are drained Beaches are the smooth sloping accumulations of for development, or when any type of land use sand and gravel along shorelines. The surface is ceases as areas become temporarily bare as con­ stable inland, but the shoreward part is subject to struction is planned for such future uses as resi­ erosion by wind and water and to deposition in pro­ dences, shopping centers, industrial sites, or subur­ tected areas. ban and rural residential subdivisions. Land being altered by filling, such as occurs in spoil dumps or 73. SANDY AREAS OTHER THAN BEACHES sanitary landfills, also is indicative of this transi­ Sandy Areas other than Beaches are composed tional phase. primarily of dunes-accumulations of sand trans­ ported by the wind. Sand accumulations most com­ 77. MIXED BARREN LAND monly are found in deserts although they also oc­ The Mixed Barren Land category is used when a cur on coastal plains, river flood plains, and deltas mixture of Barren Land features occurs and the and in periglacial environments. When such sand dominant land use occupies less than two-thirds of accumulations are encountered in tundra areas, they the area. Such a. situation arises, for example, in a are not included here but are placed in the Bare desert region where combinations of salt flats, sandy Ground Tundra category. areas, bare rock, surface extraction, and transi- 20 A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA tional activities could occur in close proximity and engender new drainage patterns with resultant in areal extent too small for each to be included at changes in plant community or erosion character­ mapping scale. Where more than one-third inter­ istics (Price, 1972). mixture of another use or uses occurs in a specific The boundaries between Tundra, Perennial Snow area, it is classified as Mixed Barren Land. Where or Ice, and Water are best determined by using the intermixed land use or uses total less than one­ images acquired in late summer. The Forest Land­ third of the specific area, the category appropriate Tundra boundary in the Arctic tends to be transi­ to the dominant type of Barren Land is applied. tional over a wide area and characterized by either incursion of forests where site improvement occurs, 8. TUNDRA as along the flood plains or river valleys, or by in­ Tundra is the term applied to the treeless regions creasing environmental severity, as on exposed dry beyond the limit of the boreal forest and above the uplands. This Forest Land-Tundra boundary is much altitudinal limit of trees in high mountain ranges. easier to delineate in alpine areas. The Barren Land­ In the United States, tundra occurs primarily in Tundra interface occurs where one or more of the Alaska, in several areas of the western high moun­ environmental parameters necessary for vegetation tain ranges, and in small isolated locations in the growth is deficient and also would be determined higher mountains of New England and northern best with late-summer imagers. New York. The timber line which separates forest Using the results of various investigations, Level and tundra in alpine regions corresponds to an arctic II categories of Tundra based primarily on what is transition zone in which trees increasingly are re­ interpretable from remote sensor image signatures stricted to the most favorable sites. are: Shrub and Brush Tundra, Herbaceous Tundra, The vegetative cover of the tundra is low, Bare Ground Tundra, Wet Tundra, and Mixed dwarfed, and often forms a complete mat. These Tundra. plant characteristics are in large part the result of 81. SHRUB AND BRUSH TUNDRA adaptation to the physical environment-one of the most extreme on Earth, where temperatures may The Shrub and Brush Tundra category consists average above freezing only 1 or 2 months out of of the various woody shrubs and brushy thickets the year, where .strong desiccating winds may occur, found in the tundra environment. These occur in where great variation in solar energy received may dense-to-open evergreen and deciduous thickets, with exist, and where permafrost is encountered almost the latter dominated by types such as the various everywhere beneath the vegetative cover. birches (Betula), alders (Alnus), or willows The number of species in the tundra flora is rela­ (Salix), as well as many types of berry plants. Low tively small compared with typical middle- and low­ evergreen shrub thickets are characterized by such latitude flora, and this number of species decreases dominant types as Empetrum and various members as the environment becomes increasingly severe of the heath family, such as Cassiope, Vaccinium, with changes of latitude and altitude. The tundra and Ledum (Viereck and Little, 1972). vegetation consists primarily of grasses, sedges, 82. HERBACEOUS TUNDRA small flowering herbs, low shrubs, lichens, and Herbaceous Tundra is composed of various sedges, mosses. The vegetative cover 'is most luxuriant near grasses, forbs, lichens, and mosses, all of which lack the boreal forest, with the ground surface usually woody stems. A wide variety of such herbaceous being completely covered. As the plant cover be­ types may be found in close proximity on the tundra. comes sparse, shrubs become fewer and more bare Sites having sufficient moisture usually are covered areas occur. Species diversity is lowest near the with a thick mat of mosses together with sedges boundaries of permanent ice and snow areas, where such as Carex and Eriophorum (cotton grass) in only isolated patches of vegetation occur on the bare almost continuous and uniform tussocks, as well as ground surface. other herbaceous forms such as types of bluegrass The vegetation of the tundra is closely associated (Po a), buttercups (Ranunculus), and lichens such with other environmental factors. Minor manmade as Cladon1:a and Cet1·a1·ia. Drier or more exposed disturbances, as well as microenvironmental changes sites usually trend toward a sparse moss-lichen mat. over short distances, can have significant effects. Minor changes in available moisture or wind protec­ 83. BARE GROUND TUNDRA tion, for example, can result in different plant asso­ The Bare Ground Tundra category is intended for ciations. Similarly, man's activity in the tundra may those tundra occurrences which are less than one- DEFINITIONS 21 third vegetated. It usually consists of sites visually Land, or Tundra, with their common boundaries dominated by considerable areas of exposed bare being distinguished most readily on late summer rock, sand, or gravel interspersed with low herbace­ imagery. ous and shrubby plants. This type of tundra is in­ The terminology and nomenclature of any sub­ dicative of the most severe environmental stress division of Perennial Snow or Ice areas are always and usually occurs poleward of the areas supporting subject to considerable debate, but a Level II break­ the more luxuriant herbaceous and shrub forms and down into categories of Perennial Snowfields and on· higher mountain ridges. The various species of Glaciers seems to be appropriate for use with remote Dryas, such as white mountain-avens, are dominant sensor data. Such a subdivision is based on surface in Arctic regions, as are the sandworts (Minu ... form and the presence or absence of features indi­ artia) and mountainheaths (Phyllodoce). Bare cating glacial flow. In addition, these forms and Ground Tundra gradually merges with one or more flow features may be related to stage of develop­ of the Barren Land categories on its more severe ment and certain periglacial or glacial processes. margin. 91. PERENNIAL SNOWFIELDS 84. WET TUNDRA Perennial Snowfields are accumulations of snow Wet Tundra is usually found in areas having little and firn that did not entirely melt during previous topographic relief. Standing water is almost always summers. Snowfields can be quite extensive and present during months when temperatures average thus representative of a regional climate, or can above the freezing level. Numerous shallow lakes be quite isolated and localized, when they are known are also common (Joint Federal-State Land Use by various terms, such as snowbanks. Planning Commission for Alaska, 1973). Perma­ The regional snowline is controlled by general frost is usually close to the surface, and various climatic conditions and closely parallels the regional patterned ground features may be evident. Sedges 32°F (0°C) isotherm for the average temperature (Car ex) such as cotton grass are characteristically of the warmest summer month. The use of the term dominant, and a few shrubby plants may occur on "line" is somewhat misleading, because the "snow­ adjacent drier sites. Rooted aquatic plants are also line" represents an irregular transitional boundary, common. Wet Tundra is delineated best on imagery which is determined at any single location by the acquired in late summer. combination of snowfall and ablation, variables which can change greatly within short distances be­ 85. MIXED TUNDRA cause of changes in local topography and slope The Mixed Tundra category is used for a mixture orientation. of the Level II Tundra occurrences where any Small isolated snowfields occurring in protected particular type occupies less than two-thirds of the locations can develop into incipient or nivation area of the mapping unit. Where more than one­ cirques, which become gradually hollowed by the third intermixture of another use or uses occurs in annual patterns of freezing and thawing, aided by a specific area, it is classified as Mixed Tundra. downslope movement of rock material. They are Where the intermixed land cover categories total less circular to semicircular and often develop ridges of than one-third of the specific area, the category ap­ mass-wasted materials called protalus ramparts at propriate to the dominant type of Tundra is applied. their downslope margins. As Flint (1957) has pointed out, "Such cirques, of course, are not in 9. PERENNIAL SNOW OR ICE themselves indication of glaciation, they indicate Certain lands have a perennial cover of either merely a frost climate." snow or ice because of a combination of environ­ Snowfields can normally be distinguished from mental factors which cause these features to sur­ the following Glacier category by their relative lack vive the summer melting season. In doing so, they of flow features. persist as relatively permanent features on the land­ scape and may be used as environmental surrogates. 92. GLACIERS Snow, firn (coarse, compacted granular snow), or ice Glacial ice originates from the compaction of snow accumulation in these areas exceeds ablation, which into firn and finally to ice under the weight of sev­ is the combined loss of snow or ice mass by evapora­ eral successive annual accumulations. Refrozen melt tion and melt-water runoff. Adjacent lands most water usually contributes to the increasing density commonly will be classed as Water, Wetland, Barren of the glacial ice mass. With sufficient thickness, 22 A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA

weight, and bulk, flow begins, and all glaciers ex­ TABLE 4.-U.S.G.S. Level I Land Use Color Code hibit evidence of present or past motion in the 1. Urban or Built-up Land __ Red (Munsell5R 6/12). form of moraines, crevasses, and so forth. 2. Agricultural Land ______Light Brown (Munsell 5YR 7/4). Where the snowline of adjacent ice-free areas 3. Rangeland ______Light Orange (MunselllOYR 9/4). extends across the glacier, it is known as the firn 4. Forest Land ------Green (Munsell lOGY 8/5). limit, which represents the dividing line between the 5. Water ______Dark Blue (Munsell lOB 7 /7). 6. Wetland ____ .., ______Light Blue (Munsell 7.5B glacier's two major zones, the zone of accumulation 8.5/3). and the zone of ablation. While glaciers normally 7. Barren Land ______Gray (Munsell N 8/0). 8. Tundra ______Green-Gray (Munsell lOG 8.5/ are recognized easily, certain glacial boundaries may 1.5). be subject to misinterpretation, even by the experi­ 9. Perennial Snow or Ice ___ White (Munsell N 10/0). enced interpreter. Flow features upglacier from the firn limit typically are obscured by fresh snow, forc­ fication by color is necessary at Level II since it ing the image interpreter to depend on secondary would be a considerable problem to select 37 differ­ information such as valley shape or seek a more ent colors which would be distinguishable at the discriminating sensor. Similarly, morainal material size of the minimum mapping unit. A numerical sys­ may cover the terminus (or snout) of the glacier tem, with the number of digits equaling the level because of ablation, making boundary determination of categorization, forms a flexible classification sys­ in that vicinity difficult. This latter problem occa­ tem that permits continuation to Levels III and IV sionally is compounded by the presence of consider­ or beyond. In addition, retaining a discrete color code able vegetation rooted in the insulating blanket of for each Level I land use or land cover category ablation moraine. permits rapid visual integration of the areas char­ Further subdivision of glacial occurrences, mainly acterized by that use or cover type. on the basis of form and topographic position, would Even though a numerical system for the Level II include: small drift glaciers (sometimes called land uses has been illustrated, such a system is not Ural-type or cirque glaciers) ; valley glaciers (also the only method of presenting Level II land use in­ called mountain or alpine glaciers) ; piedmont gla­ formation. What is proposed is the use of the modi­ ciers ; and icecaps (or ice sheets) . fied International Geographical Union World Land Other features have somewhat the surface form Use Survey color code at Level I. Alternatives to a of true glaciers, such as "rock glaciers." Since these numerical code at Level II could take the form of are composed primarily of fragmented rock mate­ graphic symbols such as dots, stipples, cross-hatch­ rial together with interstitial ice, they are classified ing, swamp or marsh symbols, or any of the great as Bare Exposed Rock. variety of such items available to the cartographer. Such a method, together with the Level I color cod­ MAP PRESENTATION ing, would allow the reader rapid visual orientation to each discrete Level II land use category but Figures 1 through 4 depict typical maps which would impede statistical inventory of the area in­ have been produced using the U.S. Geological Sur­ vey land use and land cover classification system. cluded in each land use and would be difficult to The land use and land cover maps have been pro­ subdivide further into Level III categories. duced by conventional interpretation techniques and Another alternative for land use symbolization at are typical examples of maps produced from high­ Level II is the use of an alphabetical code for each altitude color-infrared photographs. category such as "Ur," representing (Urban or In order to provide a systematic and uniform ap­ Built-up) Residential Land, or "Ac," for (Agricul­ proach to the presentation of land use and land tural) Cropland and Pasture. Such a system has the cover information in map format, a scheme of color merit of suggesting the logical name of each cate­ coding is employed (table 4). In this scheme, Level gory but also impedes interpretation and enumer­ I land uses are color coded using a modified version ation at the more detailed levels because of increased of the World Land Use Survey (International Geo­ graphical Union, 1952) color scheme. Level II land complexity of the alphabetical code. In addition, the uses can be presented using the two-digit numeral increase in length of the alphabetical code used for appropriate to the land use category, such as "21," the more detailed levels will cause placement prob­ which would signify Cropland and Pasture. The use lems as the minimum size of a mapping unit is ap­ of some type of system other than a further strati- proached. MAP PRESENTATION 23

EXPLANATION 1 Urban or built-up land 2 Agricultural land 4 Forest land 5 Water 7 Barren land

0 2 MILES

0 2 KILOMETRES

FIGURE 1.-Level I land use and land cover in an enlarged part of the northeast quarter of the Indian­ apolis, Indiana-Illinois, 1: 25·0,000 quadrangle. Area outlined in center of map corresponds to May­ wood area shown in figures 3 and 4. 24 A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA

EXPLANATION 11 Residential 12 Commercial and services 13 Industrial 14 Transportation, communications, and utilities 15 Industrial and commercial complexes 16 Mixed urban or built-up land 17 Other urban or built-up land 21 Cropland and pasture 22 Orchards, groves, vineyards, nurseries, and ornamental horticultural areas 23 Confined feeding operations 41 Deciduous forest land 51 Streams and canals 53 Reservoirs 75 Strip mines, quarries, and gravel pits 76 Transitional areas

39° 40'

0 2 MILES f-----.-L,--___J 0 2 KILOMETRES

FIGURE 2.-Level II land use and land cover in an enlarged part of the northeast quarter of the Indianapolis, Indiana­ Illinois, 1:250,000 quadrangle. Area outlined in center of map corresponds to Maywood area shown in figures 3 and 4. MAP PRESENTATION

l~_ ~~2_8 c~~~=~', EXPLANATION ' 11 Residential 12 Commercial and services 13 Industrial 14 Transportation, communications, and utilities 17 Other urban or built-up land 21 Cropland and pasture 22 Orchards, groves, vineyards, nurseries, and ornamental horticultural areas 24 Other agricultural land 41 Deciduous forest land 42 Evergreen forest land 43 Mixed forest land 51 Streams and canals 52 Lakes 53 Reservoirs 62 Nonforested wetland 2i____,_ 75 Strip mines, quarries, and gravel pits

0 Y,. MILE

0 .5 KILOMETRE

FIGURE 3.-Level II land use and land cover in a part of the Maywood, Indiana, 1:24,000 quadrangle. Level III inter­ pretations for the same area are shown in figure 4. 26 A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA

EXPLANATION 111 Single family 122 Retail trade 131 Primary processing 132 Fabrication 134 Extraction facilities 141 Highways 144 Airports 145 Communications 147 Utilities 39° 173 Waste dumps 42' 174 Urban undeveloped 30" 211 Cropland 212 Pastureland 224 Nurseries and floriculture 242 Farmsteads 753 412 10-30 percent crown cover, ~ deciduous 413 30-70 percent crown cover, deciduous 414 >70 percent crown cover, deciduous 424 > 70 percent crown cover, evergreen 432 10-30 percent crown cover, mixed 511 Streams 521 Lakes 532 Water-filled quarries 622 Mudflats 753 Sand and gravel pits (active)

0 \>2 MILE

0 .5 KILOMETRE

FIGURE 4.-Level III land use and land cover in a part of the Maywood, Indiana, 1:24,000 quadrangle. Level II in­ terpretations for the same area are shown in figure 3. SELECTED BIBLIOGRAPHY 27

SELECTED BIBLIOGRAPHY Kuchler, A. W., 1964, Potential natural vegetation of the con­ terminous United States: Amer. Geog. Soc., Spec. Pub. Anderson, James R., 1971, Land use classification schemes no. 36, 116 p. used in selected recent geographic applications of remote Marschner, F. J., 1950, Major land uses• in the United States sensing: Photogramm.Eng., v. 37, no. 4, p. 379-387. [map, scale 1:5,000,000]: U.S. Dept. of Agriculture, Agr. Anderson, James R., Hardy, Ernest E., and Roach, John T., Research Service. 1972, A land-use classification system for use with re­ National Academy of Sciences, 1970, Remote sensing with mote-sensor data: U.S. Geol. Survey Circ. 671, 16 p., refs. special referenc.e to agriculture and forestry: Washing­ Avery, T. Eugene, 1968, Interpretation of aerial photographs ton, D.C., Natl. Acad. Sci., 423 p. [2nd ed.]: Minneapolis, Burgess Pub. Co., 324 p. New York State Office of Planning Coordination, 1969, Land Barlowe, Raleigh, 1972, Land resource economics [2nd ed.]: use and natural resources inventory of New York State: Englewood Cliffs, N.J., Pre~tice-Hall, Inc., 585 p. Albany, New York State Office of Planning Coordination, Burley, Terence M., 1961, Land use or land utilization?: Prof. 67 p. Geographer, v. 13, no. 6, p. 18-20. Oosting, Henry J., 1956, The study of plant communities Clawson, Marion, and Stewart, Charles L., 1965, Land use in­ [2nd ed.]: San Francisco W. H. Freeman Co., 440 p. formation. A critical survey of U.S. statistics including Orning, George W., and Maki, Les, 1972, Land management possibilities for greater uniformity: Baltimore, Md., The information in northwest Minnesota: Minneapolis, Univ. Johns Hopkins Press for Resources for the Future, Inc., of Minn. Center for Urban Studies, Minn. Land Manage­ 402 p. ment Inf. System Study, Rept. no. 1. Colvocoresses, Alden P., 1971, Image resolution for ERTS, Pettinger, L. R., and Poulton, C. E., 1970, The application of Skylab, and Gemini/ Apollo: Photogramm. Eng., v. 38, high altitude photography for vegetation resource in­ no. 1, p. 33-36. ventories in· southeastern Arizona: Final Rept., Contract Colvocoresses, Alden P., and McEwen, Robert B., 1973, Prog­ no. N AS 9-8577, N atl. Aeronautics and Space Admin., ress in cartography, EROS program: Symposium on 147 p. Significant Results Obtained from the Earth Resources Price, Larry W., 1972, The periglacial environment, perma­ Technology Satellite-1, N atl. Aeronautics and Space frost, and man: Washington, D.C., Assoc. of Amer. Admin. Pub. SP-327, p. 887-898. Geographers, Comm. on College Geography, Resource Ellefsen, R., Swain, P. H., and Wray, J. R., 1973, Urban Paper No. 14, 88 p. land use mapping by machine processing of ERTS-1 Rosenberg, Paul, 1971, Resolution, detectability, and recog­ multispectral data: A San Francisco Bay area example: nizability: Photogramm. Eng., v. 37, no. 12, p. 1255- West Lafayette, Ind., Purdue Univ. Lab. for Applications 1258. of Remote Sensing Inf. Note 101573. Shaw, Samuel P., and Fredine, C. Gordon, 1956, Wetlands of Flint, R. F., 1957, Glacial and Pleistocene geology: New the United States: U.S. Dept. of the Interior, Fish and York, John Wiley and Sons, Inc., 553 p. Wildlife Service Circ. 39. Frey, H. Thomas, 1973, Major uses of land in the United Shelford, Victor E., 1963, The ecology of North America: States-summary for 1969: U.S. Dept. of Agriculture, Urbana, Univ. of Illinois Press, 810 p. Econ.~·esearch Service, Agr. Econ. Rept. no. 247. Gleason, Henry A., and Cronquist, Arthur, 1964, The natural Stevens, Alan R., Ogden, W. H., Wright, H. B., and Craven, geography of plants: New York, Columbia Univ. Press, C. W., 1974, Alternatives for land use/cover mapping in 420 p. the Tennessee River watershed: Amer. Cong. on Survey­ Grigg, David, 1965, The logic of regional systems: Annals ing and Mapping, Amer. Soc. of Photogramm., ann. mtg., Assoc. Amer. Geographers, v. 55, no. 3, p. 465-491. 34th, St. Louis, Mo., Mar. 10-15, 1974, p. 533-542:. Hardy, Ernest E., Belcher, Donald J., and Phillips, Elmer S., Stoddard, Lawrence A., and Smith, Arthur D., 1955, Range 1971, Land use classification with simulated satellite pho­ management [2nd ed.]: New York, McGraw-Hill Book tography: U.S. Dept. of Agriculture, Econ. Research Co., 433 p. Service, Agr. Inf. Bull., 352 p. Sweet, David C., and Wells, Terry L., 1973, Resource man­ Hardy, Ernest E., and Shelton, Ronald L., 1970, Inventory­ agement implications of ERTS-1 data to Ohio: Sympo­ ing New York's land use and natural resources: New sium on Significant Results Obtained from the Earth Re­ York's Food and Life Sciences, v. 3, no. 4, p. 4-7. sources Technology Satellite-1, Natl. Aeronautics and Space Admin. Pub. SP-327, p. 1459-1466. Hawley, Arthur J., 1973, The present and future status of Eastern North Carolina wetlands: Chapel Hill, Univ. of Thrower, Norman J. W., 1970, Land use in the Southwestern North Carolina, Water R·esources Res. Inst., Rept. no. 87. United States from Gemini and Apollo imagery (map suppl. no. 12): Annals Assoc. Amer. Geographers, v. Ingram, J. J., and Prochaska, D. D., 1972, Me·asuring com­ 60, no. 1. pleteness of coverage in the 1969 census of agriculture: Am. Stat. Assoc., Business and Econ. Sect., ann. mtg., U.S. Bureau of the Census, 1970, Areas of the United States: Montreal 1972, Proc., p. 199-215.. U.S. Dept. of Commerce, Bureau of the Census, Area Measurement Rept. GE-20, no. 1. International Geographical Union, 1952, Report of the commit­ tee on world land survey for the period 1949-1952: U.S. Congress, 1936, The Western Range: U.S. 74th Cong., Worcester, England, 23 p. 2d s.,ess., Senate Doc. 199. Joint Federal-State Land Use Planning Commission for --- 1973, The land use policy and planning assistance act: Alaska, 1973, Major ecosystems of Alaska: Anchorage, U.S. 93rd Cong., 1st sess., Senate Bill 268. Joint Federal-State Land Use Planning Comm. for Alaska, U.S. Department of Agriculture, Conservation Needs In­ map, scale 1:2,500,000, incl. text. ventory Committee, 1971, National inventory of soil and 28 A LAND USE AND LAND COVER CLASSIFICATION SYSTEM FOR USE WITH REMOTE SENSOR DATA

water conservation needs, 1967: Statistical Bull. 461, ard land use coding manual, a standard system for 211 p. identifying and coding land use activities: Washington, --- 1972, Farmland: Are we running out?: The Farm D.C., 111 p. Index, v. XI no. 12, p. 8-10. Viereck, Leslie A., and Little, Elbert L., Jr., 1972, Alaska U.S. Department of Agriculture, Soil Conservation Service, trees and shrubs: U.S. Dept. of Agriculture, Forest Serv­ 1962, Classifying rangeland for conservation and plan­ ice Handbook 410, 2:65 p. ning: U.S. Dept. of Agr. Handbook 235. Welch, Roy, 1973, Cartographic quality of ERTS-1 image: [U.S.] Executive Office of the President, Bureau of the Bud­ Symposium on Significant Results Obtained from the get, 1957, Standard industrial classification code: Wash­ Earth Resources Technology Satellite-!, Natl. Aeronau­ ington, D.C. tics and Space Admin. Pub. SP-327, p. 879-886. U.S. Geological Survey, 1973, Geological Survey reseaTCh Wooten, Hugh H., and Anderson, James R., 1957, Major uses 1973: U.S. Geol. Survey Prof. Paper 850, p. 255-258. of land in the United States~summary for 1954: U.S. U.S. Urban Renewal Administration, Housing and Home Fi­ Dept. of Agriculture, Agr. Research Service, Agr. Inf. nance Agency, and Bure•au of Public Roads, 1965, Stand- Bull. 168.